1. Field of the Invention
This invention relates to light emitting diodes (LEDs) and more particularly to an apparatus with multiple LEDs that in combination produce white light.
2. Description of the Related Art
Light emitting diodes (LEDs) are an important class of solid-state devices that convert electric energy to light. They generally comprise an active layer of semiconductor material sandwiched between two oppositely doped layers. When a bias is applied across the doped layers, holes and electrons are injected into the active layer where they recombine to generate light. Light is emitted omnidirectionally from the active layer and from all surfaces of the LED. Recent advances in LEDs (such as nitride based LEDs) have resulted in highly efficient light sources that surpass the efficiency of filament-based light sources, providing light with equal or greater brightness in relation to input power.
One disadvantage of conventional LEDs used for lighting applications is that they cannot generate white light from their active layers. One way to produce white light from conventional LEDs is to combine different wavelengths of light from different LEDs. For example, white light can be produced by combining the light from red, green and blue emitting LEDs, or combining the light from blue and yellow LEDs.
One disadvantage of this approach is that it requires the use of multiple LEDs to produce a single color of light, increasing the overall cost and complexity. In addition, the different colors of light are often generated from different types of LEDs fabricated from different material systems. Combining different LED types to form a white lamp can require costly fabrication techniques and can require complex control circuitry since each device may have different electrical requirements and may behave differently under varied operating conditions (e.g. with temperature, current or time).
More recently, the light from a single blue emitting LED has been converted to white light by surrounding the LED with a yellow phosphor, polymer or dye, with a typical phosphor being cerium-doped yttrium aluminum garnet (Ce:YAG). [See Nichia Corp. white LED, Part No. NSPW300BS, NSPW312BS, etc.; See also U.S. Pat. No. 5,959,316 to Hayden, “Multiple Encapsulation of Phosphor-LED Devices”]. The surrounding phosphor material “downconverts” the wavelength of at least some of the LED light, changing its color. For example, if a nitride-based blue emitting LED is surrounded by a yellow phosphor, some of the blue light passes through the phosphor without being changed while a substantial portion of the remaining light is downconverted to yellow. The LED will thus emit both blue and yellow light, which combine to provide a white light.
This approach has been successfully used to commercialize white LEDs for a variety of applications such as flashlights, indicator lights, display backlighting, and architectural lighting. However, conventional blue LEDs are too dim for many general lighting applications that currently make use of filament-based or fluorescent lamps. While improvements in blue LED efficiency and output power would be beneficial in increasing the light output from white LEDs, a number of other factors exist which limit the performance of such devices. For example, phosphor materials have a finite conversion efficiency, resulting in “conversion loss” since a portion of the absorbed light is not re-emitted as downconverted light. Additionally, each time a higher energy (e.g., blue) photon is converted to a lower energy (e.g., yellow) photon, light energy is lost (Stokes loss), resulting in an overall decrease in white LED efficiency. This reduction in efficiency increases as the gap between the wavelengths of the absorbed and re-emitted (downconverted) light increases. Finally, for blue LEDs to emit an output light flux sufficient for room illumination, the LED chips themselves can become very hot, causing damage the component device layers of the LED chip itself, or degrading surrounding encapsulation or downconverting media.
Another disadvantage of the above white light emitter arrangements (red+green+blue LEDs or blue LEDs combined with yellow phosphors) is that they do not produce the optimal spectral emission necessary for both high efficacy and high color rendering. Simulations of white emitters show that high efficacy and color rendering can be achieved with an output light spectrum consisting of spectrally narrow emission in the blue and red regions, with a slightly broader emission in the green region.
In the case of the red+green+blue LED lamps, the spectral emission lines of the component LEDs are typically narrow (e.g. 10-30 nm full width at half maximum (FWHM)). While it is possible to achieve fairly high values of efficacy and color rendering with this approach, wavelength ranges exist in which it is difficult to obtain high-efficiency LEDs (e.g. approximately 550 nm). As a result, it is difficult to achieve both high efficacy and high color rendering index with low manufacturing cost and high yield. This can be particularly problematic when spectral requirements call for high efficiency green LEDs, since such LEDs have only been realized in the (In, Ga, Al)N system and are typically subject to low yield and strong wavelength and emission variations with operating conditions such as drive current and temperature. While more simplified white lamps may be realized using only two LEDs emitting at complimentary colors (e.g. blue, yellow), it is exceedingly difficult to achieve high color rendering coefficients in such lamps, primarily due to the lack of any red light in the resulting spectrum.
In the case of blue LEDs combined with yellow phosphor, the resulting white light is produced without a red light source. Since the resulting light is typically deficient in one of the primary colors, lamps fabricated in this manner display poor color rendering.
The desired spectrum can be more closely achieved using a combination of a blue LED with two separate phosphors emitting in the green and red spectral regions, or using an ultra violet LED with red, green and blue phosphors. However, suitable red phosphors having high conversion efficiency and the desired excitation and emission characteristics have yet to be reported. Even if such red phosphors were available, they would be subject to significant energy (Stokes) losses due to the conversion of high energy blue or UV photons to lower energy red photons.
Patent Publication No. US 2002/0070681 A1 to Shimizu discloses an LED lamp exhibiting a spectrum closer to the desired spectrum. The lamp has a blue LED for producing a blue wavelength light, a red emitting LED, and a phosphor which is photoexcited by the emission of the blue LED to exhibit luminescence having an emission spectrum wavelength between the blue and red wavelength spectrum. The phosphor is preferably a yellow or green emitting phosphor and in the embodiments shown, the phosphor covers both the red and blue LEDs.
One of the disadvantages of the Shimizu lamp is that blue LEDs are not as efficient as other LEDs emitting at other wavelength spectrums and a limited number of phosphors are available for luminescence from a blue wavelength of light. Another disadvantage is that with red and blue LEDs placed side by side, the projected light may have an asymmetric appearance such that the light appears red on one side and blue on the other. Since phosphor particles typically must be on the order of at least a few microns diameter to achieve high conversion efficiency (i.e., much larger than the wavelength of blue or yellow light) and particles which are larger than the wavelength of light are poor scatterers, covering one or both of the LEDs with phosphor generally does not adequately scatter the LED light to combine the different wavelengths. This can be a particular problem with large area LEDs used for high power, high output.
Another disadvantage of a number of the embodiments disclosed in Shimizu is that they show blue and red LEDs placed on top of one another and then covered by the phosphor. This can result in the shorter wavelength blue light being absorbed by the component device layers (e.g., active layers, metallization layers) of the red LED device, thereby decreasing the overall efficiency of the lamp. Also, by covering the red LED with phosphor some of the phosphor particles may absorb some of the red light, which can result in a loss of efficiency because it is generally not possible to “up-convert” the absorbed red light to higher energy green light in an efficient manner.
Shimizu also discloses an optical lens as part of its lamp, with the inside surface of the lens being roughened to increase mixing of the LED light. Such approaches are generally not effective and can decrease efficiency by interfering with the purpose of the lens, which is to reduce backscattering of light at the lens/air interface and possible subsequent re-absorption within the body of the lamp or LED.
Solid-state semiconductor lasers convert electrical energy to light in much the same way as LEDs. They are structurally similar to LEDs but include mirrors on two opposing surfaces, one of which is partially transmissive. In the case of edge emitting lasers, the mirrors are on the side surfaces; the mirrors provide optical feedback so that stimulated emission can occur. This stimulated emission provides a highly collimated/coherent light source. A vertical cavity laser works much the same as an edge emitting laser but the mirrors are on the top and the bottom. It provides a similar collimated output from its top surface. Some types of solid-state lasers can be more efficient than LEDs at converting electrical current to light.